Reich's AI Policy Agenda
Reich's policy framework for the AI economy consists of five interdependent interventions. First, workers whose output became AI training data must be compensated—a recognition that the collectively generated knowledge on which AI systems depend is not free raw material but a public resource requiring fair compensation structures. Second, productivity gains from AI must be broadly distributed through universal basic income funded by taxation on AI revenues—the most direct mechanism for ensuring that the workers displaced by AI retain purchasing power. Third, antitrust enforcement must prevent the concentration of AI capability in a handful of platform companies whose monopolistic position would allow them to extract excessive rents. Fourth, governments must invest in open-source AI systems governed democratically rather than by corporate boards. Fifth, educational institutions must pivot from training symbolic producers to cultivating directorial capacity—judgment, taste, ethical reasoning. The agenda is politically contentious and practically challenging. It is also, in Reich's assessment, the minimum required to prevent the AI transition from producing extreme concentration of wealth and power.
In the AI Story
Each element of the agenda addresses a specific market failure that unregulated AI deployment produces. Training-data compensation addresses the free-riding by AI companies on human-generated output. UBI addresses the aggregate demand problem—if AI displaces workers without maintaining their purchasing power, the economy cannot sell what it produces. Antitrust addresses the platform monopolies that would control access to productive tools. Public investment addresses the undersupply of AI capability that serves public purposes rather than profit maximization. Educational reform addresses the mismatch between what the labor market requires and what the educational system provides.
The political coalition required to implement the agenda does not yet exist. The symbolic analysts who would benefit from redistribution have not yet recognized their collective interest. The technology companies that would be taxed and regulated are mobilizing to prevent it. The broader public, which would benefit from broadly distributed AI gains, lacks the technical understanding to engage with the specifics. Reich's prediction—that the professional class itself will become the constituency demanding UBI—depends on the symbolic analysts recognizing that their individual positions are deteriorating and that collective political action is the only response capable of addressing the structural causes.
The feasibility of the agenda is contested. Critics argue that AI companies will relocate to jurisdictions with lighter regulation, that UBI will reduce work incentives, that public AI investment will be captured by the same interests that capture other public resources. Reich's response is that these objections have been raised against every previous social contract and that the contracts were built despite them. The New Deal faced identical objections. So did the post-war labor compact. The objections did not prevent the contracts from being built. They shaped the negotiations, and the negotiations produced imperfect but functional agreements that distributed the gains from economic growth more broadly than unregulated markets would have.
Origin
The five points are synthesized in Chapter 10 of this volume from Reich's dispersed 2023-2026 essays and interviews. Each point has antecedents in Reich's earlier work: the training-data compensation extends his arguments about intellectual property and the digital commons; UBI extends his work on inequality and redistribution; antitrust extends his career-long focus on corporate concentration; public investment extends his advocacy for infrastructure; educational reform extends his arguments about human capital and national competitiveness.
Key Ideas
Compensation for training data. Treating human-generated output as a public resource requiring fair compensation when used commercially—new IP frameworks, data dividends, collective licensing.
UBI funded by AI taxation. Maintaining aggregate demand by ensuring displaced workers retain purchasing power—Reich predicts the professional class will demand this when they experience displacement firsthand.
Antitrust to prevent platform monopolies. Breaking up or regulating the concentration of AI capability in a handful of firms whose market power gives them excessive influence over access and pricing.
Public investment in democratically governed AI. Building open-source models and infrastructure that serve public purposes rather than profit maximization, governed by institutions accountable to citizens.
Educational reform for directors, not producers. Institutions that cultivate judgment, taste, and ethical reasoning rather than certifying the execution skills AI can replicate.
Debates & Critiques
The agenda is contested on feasibility, desirability, and political viability. Some argue it would reduce innovation and drive AI development to less-regulated jurisdictions. Others argue it does not go far enough—that platform ownership should be socialized entirely. The professional class is divided on whether to support redistribution or to rely on individual reskilling. The political window for implementing the agenda is uncertain and narrowing.
Appears in the Orange Pill Cycle
Further reading
- Robert Reich, The System: Who Rigged It, How We Fix It (2020)
- Mariana Mazzucato, Mission Economy (2021)
- Joseph Stiglitz, People, Power, and Profits (2019)
- Glen Weyl and Eric Posner, Radical Markets (2018)